-
公开(公告)号：KR100966592B1
公开(公告)日：2010-06-29
申请号：KR1020070132671
申请日：2007-12-17
Applicant: 한국전자통신연구원
CPC classification number: G06T7/593 , G06T7/85 , G06T2207/10012
Abstract: 본 발명은 카메라의 보정 방법에 관한 것으로, 보정하고자 하는 카메라로 찍은 영상이 적은 수이고 배경에 존재하는 기하 정보로서 평행 사변형 밖에 없을 때, 소실점의 대응을 이용하는 방법에 비해 정밀도가 높은 카메라 보정을 수행 할 수 있으며, 같은 직선상에 놓이지 않은 소실점이 4개 이상이 아닌 3개뿐인 경우에도 영상사이의 많은 대응점을 필요로 하는 프로젝티브 기하 복원과정이 없이도 무한 호모그래피를 구하여 카메라 보정이 가능하고, 사용자가 특별한 보정틀을 들고 다니지 않고 많은 수의 영상을 획득하지 못한 상황이라 하더라도 평행 사변형 영상을 찍은 적은 수의 영상만 있으면 카메라를 보정하여 카메라로 3차원 정보를 구하는 작업에 이용할 수 있는 이점이 있다.
내부변수 보정, 교정 호모그래피, 평행 사변형 기하 정보Abstract translation: 一种用于校准照相机的方法，包括（a）获得将投影到照相机拍摄的两幅图像上的每个平行四边形映射成矩形的二维（2D）单应性，其中2D单应性被定义为整流单应性，并且其中新的照相机 虚拟图像被定义为整流照相机，并且在两个整流摄像机之间产生新的无限单对应性，虚拟图像由整流单相图像从原始图像转换; （b）通过使用新的无限单体图，整流单应性和原始无限单体图之间的相关性来获得原始的无限单体图; 和（c）基于原始无限单色图和本征相机参数之间的相关性来获得固有的相机参数，从而校准相机。
-
公开(公告)号：KR1020100053411A
公开(公告)日：2010-05-20
申请号：KR1020090024051
申请日：2009-03-20
Applicant: 한국전자통신연구원
IPC: G06T7/00
CPC classification number: G06T7/246 , G06T7/11 , G06T7/44 , G06T2207/20021 , G06T2210/28
Abstract: PURPOSE: A multi-core multi-thread based KLT(Kanade Lucas Tomasi) feature point tracking method and apparatus are provided to obtain input data of a system needing data on two-dimensional feature points by extracting parallel real-time KLT feature points according to region and tracking the extracted feature points. CONSTITUTION: A division and allocation unit performs multiple image division and core allocation according to region(S401). An intrinsic value calculator calculates intrinsic values(S415). A feature point extractor extracts a proper number of final feature points through the arrangement of the intrinsic values(S419). A moment value calculating unit calculates moment values through the feature points and gradient value(S423). A displacement estimating unit computes estimated displacement through the moment values(S427). The displacement estimating unit presumes the final displacement(S429).
Abstract translation: 目的：提供一种基于多核多线程的KLT（Kanade Lucas Tomasi）特征点跟踪方法和装置，通过提取并行实时KLT特征点，获得需要二维特征点数据的系统的输入数据 区域并跟踪提取的特征点。 构成：划分和分配单元根据区域执行多个图像分割和核心分配（S401）。 内在值计算器计算内在值（S415）。 特征点提取器通过内在值的排列提取适当数量的最终特征点（S419）。 力矩值计算单元通过特征点和梯度值计算力矩值（S423）。 位移估计单元通过力矩值计算估计位移（S427）。 位移估计单元假设最终位移（S429）。
-
公开(公告)号：KR1020060070604A
公开(公告)日：2006-06-26
申请号：KR1020040109127
申请日：2004-12-21
Applicant: 한국전자통신연구원
IPC: H04N21/43 , G09B5/00 , G09B7/02 , H04N7/173 , H04N21/236 , H04N21/258 , H04N21/475 , H04N21/658 , H04N21/8543 , G06Q50/20
CPC classification number: H04N21/4307 , G09B5/00 , G09B7/02 , H04N7/17318 , H04N21/23614 , H04N21/25833 , H04N21/25875 , H04N21/4753 , H04N21/6581 , H04N21/8543 , G06Q50/20
Abstract: An apparatus for a SCORM-based e-learning contents service in a digital broadcasting system and a method thereof that makes it possible to implement a SCORM-based e-learning method that is a PC-based leaning method currently recognized as the standard of the e-learning industries in an MPEG-2-based digital broadcasting system. The apparatus includes a set top box for transmitting information about a kind of terminal through a return channel and then reproducing a learning TS received through a broadcasting network on a digital television (TV) receiver in accordance with received XML sync information if a user is authenticated, a learning management system for transmitting an API adaptor supportable in the set top box through the return channel in accordance with the terminal kind information transmitted from the set top box, and a digital TV transmitter for generating the learning TS by packaging learning moving-picture data, learning data and sequencing information transmitted from the learning management system, generating the XML sync information, and transmitting the learning TS and the XML sync information through the broadcasting network.
-
公开(公告)号：KR1020050046383A
公开(公告)日：2005-05-18
申请号：KR1020030080639
申请日：2003-11-14
Applicant: 한국전자통신연구원
IPC: G11B20/10
CPC classification number: G06T1/0085 , G06T1/0064 , G06T2201/0051 , G06T2201/0061 , H04N1/32208 , H04N1/32229 , H04N1/32251
Abstract: 본 발명은 디지털 데이터의 저작권을 보호하기 위한 워터마킹(watermarking) 기법에 관한 것으로, 원본프레임을 n개로 분할된 각 서브그룹으로 구분하는 단계와, 프레임의 각 픽셀 위치마다 다른 값을 가산 및 감산하는 단계와, JND(Just Noticeable Difference) 기법을 이용하여 워터마크 정보의 삽입강도(Δ) 또는 연속된 두 프레임의 평균을 적응적으로 변화시켜 워터마크 정보를 각 픽셀에 적응적으로 삽입하는 단계와, 서브그룹의 평균(m
ei , m
oi )을 산출하는 단계와, 산출되는 평균값을 이용하여 워터마크 정보를 추출하는 단계를 포함한다. 본 발명에 따른 워터마킹 기법에 의하면, 절삭, 회전, 크기변화, 프로젝션 공격에 강할 뿐만 아니라 압축, 필터링 등의 공격에도 강인하며, 압축한 후 기하학적 공격을 가하여도 삽입된 워터마크를 추출할 수 있기 때문에 보다 안정된 소유권 주장을 보장한다. 또한, 본 발명은 비디오 워터마킹 알고리즘의 요구사항인 실시간성도 완벽하게 보장할 수 있는 바 비디오 방송 워터마크 정보를 실시간으로 삽입할 수 있다는 효과가 있다.-
公开(公告)号：KR100453683B1
公开(公告)日：2004-10-20
申请号：KR1020020014114
申请日：2002-03-15
Applicant: 한국전자통신연구원
IPC: H04N1/387
CPC classification number: G06T1/0057 , G06T2201/0051 , G06T2201/0083 , H04N1/32203 , H04N1/32208 , H04N2201/3281
Abstract: Disclosed is a method for embedding and extracting a spatial domain blind watermark using a sample expansion. The sample expansion method using a sample expansion coefficient is introduced to basically improve a common disadvantage of the conventional watermarking method based on the patchwork method, for instance, the watermark is not robust against a compression attack such as JPEG. By applying the sample expansion to a sample group, the embedded watermark comes to contain a low frequency component more larger than ever, so that radio frequency component decreases as much. Thus, although the radio frequency component of the watermark is damaged by the image compression such as JPEG, it is possible to relatively reduce the damage of the watermark. Also, an adaptive watermark embedding method can be considered such that the embedded watermark value is proportional to an average value of the pixel. In order to overcome non-robustness that is a common disadvantage of the conventional methods based on the conventional patchwork method or the conventional modified patchwork method, not a single pixel but sample-expanded pixel is chosen, and there is proposed a spatial domain blind watermarking method using a watermark value that is proportional to an average value of pixel every tile.
Abstract translation: 公开了一种使用样本扩展来嵌入和提取空间域盲水印的方法。 引入使用样本展开系数的样本展开方法基本上改进了基于拼凑方法的传统水印方法的普遍缺点，例如，水印对于诸如JPEG的压缩攻击不具有鲁棒性。 通过将样本扩展应用于样本组，嵌入的水印会包含比以往更大的低频分量，使射频分量减少得更多。 因此，虽然水印的射频分量被诸如JPEG的图像压缩损坏，但可以相对减少水印的损害。 而且，可以考虑自适应水印嵌入方法，使得嵌入的水印值与像素的平均值成比例。 为了克服基于传统拼接方法或传统修改拼接方法的常规方法的普遍缺点的非鲁棒性，选择不是单个像素而是样本膨胀像素，并且提出了空间域盲水印 方法使用与每个像素的像素的平均值成比例的水印值。
-
Patent Agency Ranking
公开(公告)号：KR100375823B1
公开(公告)日：2003-03-15
申请号：KR1020000083017
申请日：2000-12-27
Applicant: 한국전자통신연구원
IPC: H03M13/41
CPC classification number: H03M13/6362 , H03M13/395 , H03M13/3961 , H03M13/41 , H03M13/4107
Abstract: A structure and a method for depuncturing an input bit stream being input to a Viterbi decoder when the Viterbi decoder is designed by using a Radix-4 branch metric calculator in a method for designing the Viterbi decoder that decodes a punctured code at a high-speed, are disclosed. A depuncture structure for Radix-4 branch metric calculation in a high-speed Viterbi decoder includes four FIFOs, four multiplexers, and one Radix-4 branch metric calculator. Two input bit streams of I and Q are connected to two upper FIFOs and two lower FIFOs. An output terminal of FIFO is connected to upper and lower multiplexers of the next stage. One output terminal of each multiplexer is connected to Radix-4 branch metric calculator. As a result, Radix-4 branch metric calculation can be achieved by using the same clock as a clock speed of the input I and Q bit streams. This structure and this method can be applied to a depuncturing process for Radix-4 branch metric calculation of all punctured codes derived from ½code.
Abstract translation: 一种用于在设计维特比解码器的方法中通过使用基数-4分支度量计算器来设计维特比解码器时输入到维特比解码器的结构和方法，所述维特比解码器以高速解码穿孔码 ，被披露。 用于高速维特比解码器中基数-4分支度量计算的解套结构包括四个FIFO，四个多路复用器和一个基数-4分支度量计算器。 I和Q的两个输入比特流连接到两个较高的FIFO和两个较低的FIFO。 FIFO的输出端连接到下一级的上下多路复用器。 每个多路复用器的一个输出端连接到基数-4分支度量计算器。 结果，基数-4分支度量计算可以通过使用与输入I和Q比特流的时钟速度相同的时钟来实现。 该结构和该方法可以应用于从＆frac12; code派生的所有收缩代码的基数-4分支度量计算的收缩处理过程。
-
公开(公告)号：KR100298212B1
公开(公告)日：2001-08-07
申请号：KR1019970069682
申请日：1997-12-17
IPC: G06F9/46
Abstract: PURPOSE: A method for scheduling tasks in real time for solving problems of a priority reversal is provided to increase the number of real-time tasks by solving problems of a priority reversal generated from the arrival of a new real-time task, and by preventing execution orders of tasks determined in an existing system from being influenced by the arrival of new tasks. CONSTITUTION: An execution of a real-time task in execution is temporarily discontinued in case that a new real-time task is arrived in a system(201). A time to spare of the task in a real-time task table is estimated(202). After that the time is estimated, it is judged whether the time is shorter than the least time to spare among real-time tasks(203). In case that the time of the new real-time task is shorter than the least time, the task temporarily discontinued is inserted into a special table for preoccupied tasks(204). A real-time task newly arrived is executed(205). In case that the time to spare of the new real-time task is compared with the least time, the new real-time task is inserted into the real-time task table(206). The task temporarily discontinued is executed again(207).
-
公开(公告)号：KR1020000032772A
公开(公告)日：2000-06-15
申请号：KR1019980049332
申请日：1998-11-17
Applicant: 한국전자통신연구원
IPC: G06F19/00
Abstract: PURPOSE: A space index composition method for a similarity search is provided to estimate an overall search cost including a space index search cost and a candidate object access cost by analyzing characteristics of a time series database in advance, and extract a dimension number of the space index for minimizing the overall cost. CONSTITUTION: A space index composition method comprises a 1st step of obtaining standard deviation of discrete fourier transform(DFT) coefficients on a time series database, arranging the standard deviation and analyzing the characteristics of the time series database in advance, a 2nd step of assuming generation of multi dimensional space index on all the possible cases made by an arbitrary DFT coefficient by referring to the characteristics, and then calculating a search cost of a query set on the space index, and a 3rd step of selecting DFT coefficients, applied to a space index minimizing the search cost, as an index entry of the multi dimensional space index.
Abstract translation: 目的：提供一种用于相似性搜索的空间索引组合方法，用于通过分析时间序列数据库的特征来预先估计包括空间索引搜索成本和候选对象访问成本的总体搜索成本，并提取空间索引维数 索引，以最大限度地降低整体成本。 构成：空间索引组合方法包括在时间序列数据库上获得离散傅里叶变换（DFT）系数的标准偏差的第一步骤，预先设置标准偏差和分析时间序列数据库的特性，第二步假设 通过参考特征对由任意DFT系数产生的所有可能情况生成多维空间索引，然后计算空间索引上的查询集合的搜索成本，以及第三步选择应用于 空间索引使搜索成本最小化，作为多维空间索引的索引条目。
-
公开(公告)号：KR101892093B1
公开(公告)日：2018-08-28
申请号：KR1020160148286
申请日：2016-11-08
Applicant: 한국전자통신연구원
IPC: H04N13/30
Abstract: 사용자의지시위치추정장치및 방법이개시된다. 본발명에따른사용자의지시위치추정방법은, 몰입형디스플레이의스크린을시청하는한 명이상의사용자가포함된다시점깊이영상을입력받는단계, 입력받은상기다시점깊이영상을분석하여, 상기사용자의머리영역및 손영역을검출하는단계, 상기머리영역의 3차원좌표및 상기손 영역의 3차원좌표를기반으로, 상기사용자의시선추적선을생성하는단계, 그리고상기시선추적선을이용하여, 상기스크린상에서상기사용자의시선에상응하는지시위치를예측하는단계를포함한다.
-
公开(公告)号：KR1020170028678A
公开(公告)日：2017-03-14
申请号：KR1020150125610
申请日：2015-09-04
Applicant: 한국전자통신연구원
Abstract: 착용형디스플레이장치에서의초점자동조절장치및 방법이개시된다. 본발명의일실시예에따른착용형디스플레이장치에서의초점자동조절장치는비접촉식거리측정장치를이용하여착용형디스플레이장치를착용한사용자의얼굴표면과상기착용형디스플레이장치에상응하는렌즈와의거리및 상기렌즈와패널사이의거리를측정하는거리측정부; 상기사용자의얼굴표면과상기렌즈사이의거리와기준값의차이에기반하여상기기준값에상응하는초점거리를유지하기위한상기렌즈의위치를계산하는렌즈위치계산부; 및계산값에기반하여상기렌즈를이동시켜상기사용자와상기렌즈사이의거리와상기렌즈와상기패널사이의거리를조절하여상기초점거리를유지시키는초점거리유지부를포함한다.
-
-
-
-
-
-
-
-
-
Search Results
72
records
(took 0.024 seconds)
